Chapter 15 Silica and Alumina Nanophases: Natural Processes …homepages.see.leeds.ac.uk/~earlgb/Publications/Tobler et al New... · Silica and Alumina Nanophases: Natural Processes

Chapter 15Silica and Alumina Nanophases: NaturalProcesses and Industrial Applications

Dominique J. Tobler, Tomasz M. Stawski, and Liane G. Benning

15.1 Introduction

Silicon (Si), aluminium (Al) and oxygen (O) are the three most abundant elementsin the Earth’s crust. The oxides these elements form constitute the main buildingblock for silicate minerals, and they are most often the end member oxides in rockweathering processes. Due to their ubiquity, the variety of polymorphs they form,each unique in structure, property and reactivity, they are among the most usefulnatural materials.

Silica, usually denoted as SiO2, can crystallise in many different crystal systems,forming a variety of polymorphs with stabilities governed by variable pressureand temperature conditions (Heaney et al. 1994). Among these, quartz is the mostcommon and most stable form of SiO2. When dissolved, the amount of silicain natural water is controlled by physicochemical parameters that are linked to

D.J. Tobler (�)Nano-Science Center, Department of Chemistry, University of Copenhagen, Copenhagen,Denmarke-mail: [email protected]

T.M. StawskiGerman Research Centre for Geosciences, GFZ, D-14473 Potsdam, Germany

School of Earth and Environment, University of Leeds, LS2 9JT Leeds, UK

L.G. BenningGerman Research Center for Geosciences, GFZ, Interface Geochemistry Section, 14473Potsdam, Germany

Department of Earth Sciences, Free University of Berlin, 12249 Berlin, Germany

School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK

© Springer International Publishing Switzerland 2017A.E.S. Van Driessche et al. (eds.), New Perspectives on Mineral Nucleation and Growth,DOI 10.1007/978-3-319-45669-0_15

293

mailto:[email protected]

294 D.J. Tobler et al.

water–rock interactions (e.g. weathering) or biomineral growth (e.g. diatoms orradiolarians). These can vary significantly between different natural environments,e.g. 100 ppb in seawater to >1000 ppm in near-boiling pressurised geothermalwaters. At Earth surface conditions, when SiO2 levels become supersaturated, silicaprecipitates as an amorphous phase. Such amorphous silica, often referred to asopal-A, has a local structural coherence smaller than 20 Å, and its solubility, density,hardness and composition (e.g. water content can vary between 1 and 15 %) arehighly dependent on precipitation conditions (Perry and Keeling-Tucker 2000).Yet, over time opal-A is unstable and gradually transforms into more crystallinephases and eventually to quartz. The time scale for this transformation process ishighly dependent on the physicochemical conditions (i.e. temperature, pressure, pHconditions) (Williams and Crerar 1985) and can take tens of thousands of years(Herdianita et al. 2000).

The alumina system (Al2O3–H2O) is notoriously more complicated, with dif-ferent polymorphs forming under different temperature and pressure conditionsbut also at different solution pH. Aqueous Al3C oxy-species (i.e. the aluminates)form as products of aqueous dissolution of aluminosilicates (which make up �75 %of the Earth’s crust) or of simpler Al(OH)3, AlOOH and Al2O3 phases, with thehighest dissolution rates occurring at acidic and alkaline conditions. In solutionssupersaturated with regard to aluminate, typically only Al(OH)3 and AlOOHphases precipitate, whereas Al2O3 or complex aluminosilicate phases are kineticallyinhibited and usually only form at higher temperatures. In the pure alumina system,about half a dozen of metastable oxide and oxyhydroxide polymorphs are known toexist, and these only transform at >1000 ıC to the thermodynamically stable trigonalcorundum (’-Al2O3). For a full review, see Levin and Brandon 1998 and referencestherein.

In both the Si and Al systems, the formation of amorphous and (micro)crystallinephases is preceded by hydrolysis and condensation reactions leading to variousmono-, di- or polynuclear species (Iler 1979; Brinker and Scherer 1990). In the silicasystem, solution speciation (Fig. 15.1a) and the type and abundance of polynuclearsilica species do not affect polymorph selection, although they are believed toinfluence the formation kinetics (Iler 1979; Rothbaum and Rohde 1979; Beltonet al. 2012). In contrast, the versatility of hydrolysed alumina species (Fig. 15.1b)and the polynuclear complexes it forms, most likely affects alumina polymorphselection (Wefers and Misra 1987; Casey 2006; Sipos 2009). In this chapter, wereview the current knowledge on the mechanisms and kinetics of polymerisationand precipitation for these two systems. For simplicity, they will be discussedseparately, first focusing on the silica system, considering both processes in naturalwaters and industrial applications. For the alumina system, we will take into accountthe amphoteric character of the solid phases, and the precipitation process will bediscussed separately for acidic and alkaline solutions.

15 Silica and Alumina Nanophases: Natural Processes and Industrial Applications 295

B 100Al3(OH)4

Al2(OH)2

AlOH

Al3+5+

-

Al(OH)21+

Al(OH)4Al(OH)30

2+

4+

90

80

70

60

50

40

30

20

10

00

pH2 4 6 8 10 12 14

% s

peci

es100A90

80

70

60

50

40

30

20

10

07 8 9 10 11

mo

l% S

iO2

pH

Si(OH)4 SiO(OH)3

Si2O2(OH)5

Si4O8(OH)44-

Si4O6(OH)62-

Si3O5(OH)53-

-

-

Fig. 15.1 Partial speciation diagrams of dissolved silica (a) and alumina (b) as a function of pH at25 ıC and 1 atm (Printed with permission from Elsevier from Dietzel (2000) for silica and Paniaset al. (2001) for alumina)

15.2 Silica in Natural Waters

Silica polymerisation and the subsequent precipitation of silica occur in many mod-ern terrestrial environments (e.g. hot springs, brines, marine sediments, rivers), andthese processes are critical to a variety of processes including sinter formation, silicadiagenesis and biosilicification (e.g. diatoms, sponges and plants) (Pancost et al.2005; Perry and Keeling-Tucker 2000; Tobler et al. 2008), as well as the transportand fate of nutrients and contaminants, with silica nanoparticles acting as adsorbentsor mobile carriers (Ryan and Elimelech 1996). Silica also played an important rolein ancient geological settings; for example, most Archean fossils were preservedin silica cherts (Westall and Walsh 2000). Furthermore, such processes were likelyalso crucial in the formation of silica-rich deposits on Mars (Squyres et al. 2008) oron other extraterrestrial bodies with the recent discovery of silica-rich nanoparticlesejected from Enceladus, Saturn’s sixth largest moon, pointing towards subsurface


hydrothermal processes (Hsu et al. 2015). The formation of silica in natural waterscan also be a nuisance to industrial facilities and equipment, for example, duringgeothermal power extractions or in water treatment systems, where silica scalingand fouling can greatly reduce performance (Gunnarsson and Arnorsson 2005).

15.2.1 Silica Chemistry

In most natural waters, the silicate ion is usually present as orthosilicic acid,Si(OH)4, a weakly acidic molecule (pKa 9.8), with the Si atom tetrahedrallycoordinated to four hydroxyl groups. Orthosilicic acid remains stable as long asits concentration is below the solubility of amorphous hydrated silica (�110–120 ppm at 25 ıC; Gunnarsson and Arnorsson 2000). Silica solubility is affectedby several parameters, including pH, SiO2 concentration, temperature, pressureand the presence of other ions, small molecules and polymers (Alexander 1954;Iler 1979; Rothbaum and Rohde 1979; Marshall and Chen 1982; Perry andKeeling-Tucker 2000; Gunnarsson and Arnorsson 2005; Patwardhan 2011). Atthe near-neutral pH of many natural waters, silica solubility is at its minimum,while it rapidly increases at pH values above 9, with orthosilicic acid becomingincreasingly ionised (Fig. 15.1a). In acidic waters (pH < 3), the silica solubility isonly slightly increased and silica polymerisation rates are at their lowest. Despitesome discrepancy between reported solubility data, SiO2 solubility increases withincreasing temperature and/or pressure, while it generally decreases with an increasein ionic strength (IS). Naturally, IS favours the dissociation of acids, which inturn may enhance SiO2 solubility (Fig. 15.1a), but likely this will only affect thesolubility at pH close to the pKa. The lack of such data hinders a more thoroughdiscussion on IS effects. The effects of inorganic and organic additives on silicasolubility and polymerisation are highly variable, and the extent and the pathwaysin which they affect silica formation are still poorly quantified, but good insightscan be found in Iler (1979) and Patwardhan (2011).

In solutions supersaturated with respect to amorphous silica, orthosilicic acidundergoes condensation reactions via the coalescence of orthosilicic acid moleculesand the concurrent release of water:

Si.OH/4 C Si.OH/4 $ .HO/3Si –O–Si.OH/3 C H2O

Further oligomerisation leads to the formation of trimers, tetramers, cyclic speciesand other polynuclear entities, thereby maximising the number of Si–O–Si bonds.Under most experimental conditions, the orthosilicic acid remains the dominantdissolved silica species (Icopini et al. 2005), while the composition and abundanceof oligomers are primarily governed by pH and SiO2 concentration but also temper-ature, ionic strength and the presence of other cations (Belton et al. 2012). Thesepolynuclear species undergo condensation and dissolution processes up to a pointwhere a critical cluster size is reached, and cluster growth becomes more likely than


cluster dissolution. This is also termed the nucleation barrier. Once reached, it leadsto the formation of three-dimensional, internally condensed and highly hydratedsilica nuclei, a few nanometres in size (Iler 1979). These nuclei can grow throughmonomer addition and/or through particle aggregation/coalescence (discussed inSect. 15.2.2) and interact with the biochemical environment in which silica willbe deposited. Overall, silica precipitates forming in surface waters are mostlyamorphous (i.e. opal-A) and not affected by oligomer type or abundance (contrary tothe alumina system, see Sect. 15.4). However, in some studies, silica precipitationrate laws have been inferred as being a consequence of polymerisation of certainoligomers (e.g. a fourth-order reaction corresponding to tetramer condensation)(Rothbaum and Rohde 1979). This is further discussed in the next section wherethe current understanding of silica formation mechanisms and kinetics is presented.

15.2.2 Silica Formation

The mechanisms of silica polymerisation and silica nanoparticle formation innatural systems have been widely investigated in both laboratory (Alexander 1954;Iler 1979; Carroll et al. 1998, 2007; Tobler et al. 2009; Tobler and Benning 2013;Kley et al. 2014) and field settings (Carroll et al. 1998; Mountain et al. 2003;Tobler et al. 2008; Meier et al. 2014), as well as through computational approaches(Noguera et al. 2015). Silica polymerisation and silica nanoparticle formation aregenerally described by a three-stage process where (1) silica polymerisation andnucleation of silica nanospheres are followed by (2) particle growth via accretion ofsilica monomers/oligomers and/or by Ostwald ripening and (3) particle aggregation.

In most natural waters (i.e. pH > 3), silica nanoparticles carry a negative charge,which becomes more negative with an increase in solution pH and/or ionic strength(Kobayashi et al. 2005; Barisik et al. 2014). This is explained by the deprotonationof surface silane groups as pH increases and a decrease in surface HC concentrationwith increasing ionic strength. Barisik et al. further observed that for particles withdiameters < 100 nm, the negative surface charge increases with a decrease in particlesize (Barisik et al. 2014). The higher the negative charge of the silica nanoparticles,the more likely it is that they remain in suspension (because of the electrostaticrepulsion) and can get transported with the water flow. However, in most naturalwaters, silica nanoparticles are not stable within the polymerising solution andthey aggregate into dense 3D networks. This is mostly due to IS effects, i.e. thepresence of charged species (e.g. coagulating cations such as Fe2C/3C or Al3C) orof “sticky” organic molecules (e.g. polysaccharides) that reduce and even neutralisethe negative surface charge (Iler 1979; Perry and Keeling-Tucker 2000; Benninget al. 2005).

Early studies estimated the size of the first silica nanoparticles forming withinthe polymerising solution to be around 2–3 nm (Iler 1979). This was confirmedand visualised through atomic force microscopy analyses (Conrad et al. 2007).More recently, advances in scattering methods have made it possible to follow


Fig. 15.2 (a) Initial steps of silica nanoparticle formation from supersaturated silica solutionsmonitored by SAXS (640 ppm SiO2, IS D 0.22, pH D 7). Homogeneous nucleation followed byinitial fast particle growth concomitant with the decrease in molybdate-reactive silica, [SiO2(aq)](modified after (Tobler et al. 2009). (b) Silica nanoparticle size distribution in polymerisingsolutions for two different ageing times (1600 ppm SiO2, IS D 0.22, modified after Tobler et al.(2009)). (c) Decrease in monomeric silica and (d) concomitant growth of silica nanoparticles insolutions with varying initial silica concentrations as indicated (pH D 7; modified after Kley et al.(2014))

in situ and in a time-resolved manner the formation of silica nanoparticles froma polymerising solution. In Tobler et al. (2009), we combined synchrotron radiationsmall-angle X-ray scattering (SAXS) (Stawski and Benning 2013) and dynamiclight scattering (DLS) flow-through experiments with (cryo)electron microscopyand monitored in situ the formation of silica nanoparticles from polymerisingsolutions where supersaturation was induced by a pH drop from about 12 to 7.These particles grew from 3 nm to approximately 8 nm, which was mirrored bya fast decrease in the orthosilicic acid concentration in solution (Fig. 15.2a, b).The size of the silica nanoparticles after 3h of polymerisation was not greatlyaffected by the tested silica concentrations (640 and 1600 ppm SiO2) and ionicstrengths (IS D 0.02–0.22), while particle aggregation initiated considerably earlierat higher SiO2 concentrations and IS. Kley et al. (2014) followed a similar approach,employing both static and dynamic light scattering to monitor silica nanoparticleformation in pH-induced silica polymerising solutions at a wider range of SiO2

concentrations (350–3000 ppm) and at two pH regimes (7 and 8). They showed that


the particle density increased and the final particle size decreased with increasingsilica content (Fig. 15.2c, d) and pH, as expected from classical nucleation andgrowth theory. Similar to the work reported in Tobler et al. (2009), Kley et al.found that particle aggregation quickly set in at high SiO2 concentrations and thatat pH 7 aggregation was more enhanced than at pH 8 (Kley et al. 2014). Theyexplained this as a consequence of an increase in negative surface charge on silicaparticles with increasing pH and with decreasing size and suggested that this leadsto repulsive forces, inhibiting particle aggregation (Barisik et al. 2014; Kley et al.2014). In an attempt to better mimic natural processes, where silica polymerisationand silica nanoparticle formation are the result of cooling (often extremely fast) of ahigh-temperature, silica-supersaturated near-neutral fluid to ambient (geothermal)or low temperatures (deep sea), we employed an approach similar to that usedin Tobler et al. (2009) but used a high-temperature flow-through experimentalsystem to monitor silica nanoparticle formation in temperature-induced silicapolymerising solutions (Tobler and Benning 2013). Our results revealed that the rateat which silica nanoparticles formed was substantially lower (around 50 %) whenpolymerisation was induced by fast cooling as opposed to pH change (both systemshaving identical initial solution composition, i.e. [SiO2]). This was evidenced bythe occurrence of a lag time in the onset of particle growth, the formation of largercritical nuclei and the absence of particle aggregation in the temperature-inducedexperiments. This retardation in silica formation is explained by the differences intime to establish supersaturation: the radical change in pH (<30 s) imposed a fasterattainment of supersaturated conditions compared to the slower (2–3 min) and moregradual change in T.

Although there are some differences in measured particle sizes and particlegrowth profiles between these three in situ studies (Fig. 15.2), they do howeveragree in the mechanistic interpretation of silica nanoparticle formation with silicapolymerisation leading to the formation of condensed critical nuclei, which growby addition of monomers according to classical growth theory up to the point wheresilica solubility is reached and/or particle aggregation sets in (Fig. 15.2a). Ostwaldripening has been suggested to be an important growth mechanism during the latestages of silica particle growth (Iler 1979; Perry and Keeling-Tucker 2000); howeverthis process could not be identified at the conditions tested in these studies. Ostwaldripening seems particularly unlikely at higher SiO2 concentrations (>1000 ppmSiO2), where particle aggregation quickly follows particle growth.

Several studies have observed the occurrence of a lag time (i.e. induction period),where no silica polymerisation and particles are detected (i.e. no significant X-ray/light scattering above background), although at the start of these experiments,the experimental solutions were already at maximum supersaturation (Rothbaumand Rohde 1979; Tobler and Benning 2013; Kley et al. 2014). The length of thisinduction period is controlled by the same factors that determine silica solubility(i.e. T, pH, IS and [SiO2]), and it decreases with an increase in silica supersaturation(Iler 1979; Rothbaum and Rohde 1979; Gunnarsson and Arnorsson 2005; Conradet al. 2007). It is generally accepted that induction periods, recently also described


as supersaturation plateaus (Noguera et al. 2015), represent the time during whichlarger polymeric cluster structures and/or particles form and dissolve again (Perryand Keeling-Tucker 2000), in agreement with classical nucleation theory. Such largestructures could be similar to the pre-nucleation clusters suggested for other mineralsystems (e.g. CaCO3; De Yoreo et al. 2017, Chap. 1). In both the silica and calciumcarbonate systems, particle nucleation likely occurs once such cluster structuresbecome dense enough to form the first solid particle.

This shows that despite the extensive research carried out in this field, thereare still a few gaps in the molecular-level mechanistic understanding of thenucleation and growth of primary silica particles in natural aqueous solution. Thisis further illustrated by the variety of kinetic models that have been derived frommeasurements of the time-dependent decrease in orthosilicic acid concentrationor silica nanoparticle growth profiles, with reaction orders ranging between 1and 6 (Tobler et al. 2009; Tobler and Benning 2013; Kley et al. 2014; Nogueraet al. 2015) and with varying dependencies on, for example, pH, [SiO2], particlesurface area or oligomer type. Due to the massive interest of industry in theproduction of monodisperse silica nanoparticle, the bulk of the research has focusedon the mechanistic and kinetic understanding of silica nanoparticles synthesisedusing organosilanes as a precursor (see Sect. 15.2.3). However, differences in theformation process may prevent the resulting models from being applied to inorganicnatural systems.

15.2.3 Silica for Industrial Applications

Silica nanoparticles are highly desirable materials for industrial applications (e.g.electronics, biotechnology, catalysis, water purification and chromatography)because of their specific structural properties (e.g. swelling capacity, strength,durability, thermal stability, dielectric properties) (Ab Rahman and Padavettan2012; Hench and West 1990; Bagwe et al. 2006; Wang et al. 2008). In naturalwaters, silica is formed from an inorganic building block, i.e. orthosilicic acid,while most synthetic silica nanoparticles are produced through the hydrolysisand condensation of alkoxysilanes, for example, tetraethoxysilane (TEOS) andtetramethoxysilane (TMOS). Highly monodisperse, spherical silica particles havebeen obtained through the so-called Stöber method (the base catalysed hydrolysisand condensation of alkoxysilanes in low-molecular-weight alcohols) (Stöber et al.1968), but other processes such as the diffusion of alkoxysilanes into a stirredaqueous solution of lysine have also been investigated (Yokoi et al. 2006). Inaqueous solutions with acidic or basic catalysts, alkoxyl groups are hydrolysed toform silanol groups and alcohol. The condensation between silanol groups (Si–OH)or between silanol and alkoxides (Si–OR) group creates siloxane bridges (Si–O–Si).These processes are affected by the type of alkoxysilane, the nature and compositionof solvent and catalyst, temperature, pH, as well as reactant mixing procedures (AbRahman and Padavettan 2012; Besselink et al. 2013), which can be tuned to

http://dx.doi.org/10.1007/978-3-319-45669-0_1


produce specific particle sizes and distributions. Silica nanoparticles prepared withalkoxysilanes have much smoother surfaces compared to silica particles formedby the condensation of orthosilicic acid in water. This is argued to be due to thereversible exchange of alkoxyl (Si–OR) and hydroxyl (Si–OH) groups, therebycreating a relaxation mechanism for further smoothening, while condensation oforthosilicic acid produces fixed Si�O�Si bonds (Carcouet et al. 2014).

The formation conditions for industrial silica nanoparticles greatly differ fromthose in nature, thus the reaction kinetics and mechanisms, to form highly size-controlled, smooth and monodispersed silica nanoparticles during the hydrolysisand condensation of alkoxysilanes, naturally deviate from those that occur in nature(Sect. 15.2.2). However, there are quite a few similarities discussed below. Scatter-ing and spectroscopy techniques combined with (cryogenic) electron microscopyand molecular modelling have provided evidence that silica polymers with an openfractal structure are the early species, which, once sufficient levels of supersaturationare reached, transform into a dense sphere, with a size of approximately 2 nm(Boukari et al. 2000; Green et al. 2003; Fouilloux et al. 2011; Carcouet et al.2014). However, there is still debate on the mechanistic steps that lead to themonodispersed mature particles formed during the hydrolysis and condensation ofalkoxysilanes, possibly also because of the large number of parameters that impactthese processes (i.e. type and concentration of alkoxysilane, solvent and catalyst,mixing procedure, etc.). Similar to silica nanoparticle formation in natural systems,growth by monomer addition (classical growth theory) has been a suggested mech-anism (Matsoukas and Gulari 1989). Several studies have argued, however, that theobserved particle growth trends could also be explained by particle aggregation(Bogush and Zukoski 1991). In the monomer addition model, nucleation occursin the initial stages leading to the formation of a set number of nuclei, which thengrow by the addition of hydrolysed monomers. On the other hand, in the aggregationmodel, nucleation occurs continuously throughout the reaction, producing nuclei ofa certain size, which first aggregate to form larger particles and then grow by furtheraddition of the primary nuclei. A recent study combining high-resolution cryogenictransmission electron microscopy (TEM) with 29Si magic angle spinning (MAS)nuclear magnetic resonance (NMR) gives convincing evidence of the formation ofmonodisperse silica nanoparticles through the aggregation of uniform 2.3 nm sizedprimary particles to larger (>10 nm) units (Fig. 15.3) (Carcouet et al. 2014). Theselarger units grow by further addition of primary particles, which collapse uponassociation from 2.3 nm in solution to about 1.3 nm when bound to an aggregate(associated with the expulsion of water). They further showed that free primaryparticles were present at all stages of particle growth (Fig. 15.3), which demonstratesthat nucleation occurs throughout the reaction. Further support for the aggregationmodel comes from a seeded growth experiment (Bogush and Zukoski 1991),where the decrease in orthosilicic acid concentration was unaffected by the numberdensity of particles, which suggested that particle growth did not occur via directcondensation of monomers onto the particle surface. In the aggregation model,the fast initial decrease in orthosilicic acid concentration (i.e. fast consumption ofmonomers once the reaction is induced) is explained by the formation of ionised,


Fig. 15.3 Schematic representation of silica nanoparticle formation mechanism modified afterCarcouet et al. (2014). Initial open structures densify to primary particles (Stage I; T1) that collapseupon association into assemblies (Stage II; T2, T3). After association, no further new assembliesare formed, and growth occurs via the addition of primary particles (Stage III; T4, arrows point ata free primary particle)

open structures of low density that collapse to form primary particles. This leads toa steady increase in particle density up to a point where these primary particles startto aggregate into more stable assemblies, leading to a sharp decrease in particledensity. Following formation of these assemblies, the particle number remainsstable, and the decrease in orthosilicic acid concentration proceeds at a much lowerrate (producing new primary particles for growth through aggregation). Similarly,Fouilloux et al. reported an initial fast increase in the particle number density upto a plateau, where particle numbers remained constant, but particle growth stilloccurred (Fouilloux et al. 2011). They argued that this result supports a mechanismwith two consecutive phases: (a) nucleation and growth and then followed by (b)phase densification. However, as pointed out by Carcouet et al., it could also be thatprimary particles with sizes lower than 3 nm were present but simply not visible bythe employed techniques of these previous studies (Carcouet et al. 2014). Also, theusually observed plateau in particle number density occurring during the particlegrowth process can be easily explained with the aggregation processes describedabove.


15.3 Alumina in Natural Waters and in Aqueous IndustrialEnvironments

Aluminium is contained in more than 250 different minerals, most noteworthy inthe form of common aluminosilicates (e.g. feldspars or clays) and the economicallyimportant bauxite ore (mixed Al oxides and hydroxides). Rock weathering and soilformation processes are to a large extent driven by the interaction of fluids withrocks leading to the dissolution of these various phases containing Al. Similar tosilica nanoparticles, colloids in the alumina–water system control the cycling ofvarious species in natural waters by introducing a charged solid-liquid interface towhich a multitude of inorganic (e.g. phosphate, arsenic, lead, iron, chromium) andorganic (e.g. humic acid) compounds can adsorb (Kasprzyk-Hordern 2004). Thisway alumina colloids are involved in cycling of anthropogenic pollutants in soils andnatural waters and have many applications in water treatment (Hu et al. 2006). Al-containing compounds and their processing are also essential for various industries,for example, metallurgy and welding, ceramic production, catalysis, electronics,pharmaceutics, cosmetics, fire prevention, paint production, etc. (Wefers and Misra1987; Levin and Brandon 1998).

The main difference between the alumina–water and silica–water systems isthat in the alumina system, there are a multitude of possible Al-oxyhydroxidephases that can precipitate from solution. Typically, amorphous colloids form at thebeginning and then transform directly in solution to crystalline polymorphs (Wefersand Misra 1987). It is generally accepted for the alumina system that the reactionpathways and the products strongly depend on the Al speciation in solution. Thevarious aqueous Al3C oxy-species (i.e. the aluminates) form through dissolutionof amphoteric Al phases under acidic and alkaline conditions, as illustrated inFig. 15.1b. In natural waters, the precipitation of Al phases usually occurs throughthe neutralisation of Al-rich acidic solutions (pH > 3–4), while in industrial settings,the neutralisation is most often from alkaline liquors (pH < 9–10, i.e. the Bayerprocess used for processing of bauxite ores). Unsurprisingly, due to the geochemicaland industrial importance of the alumina system, a substantial body of literature onprecipitation from solution as well as speciation of aqueous entities exists (e.g. seethe reviews by Wefers and Misra (1987), Swaddle et al. (1994), Casey (2006), Sipos(2009) and references therein). Below we present a brief summary of the currentknowledge.

15.3.1 Solid Phases in the Alumina–Water System

The water–alumina system is characterised by a plethora of polymorphs, and thusit is inevitable to consider first the possible solid products of precipitation fromthe aluminate solutions. Excluding pure Al2O3 phases (since they typically do not


precipitate from solution), these include (Wefers and Misra 1987; Ruan et al. 2001;Kloprogge et al. 2006; Demichelis et al. 2009; Smith et al. 2013):

• Colloidal amorphous and gelatinous hydrated alumina(s) formed through con-densation of aluminate complexes (Bale and Schmidt 1959; Petz 1968; Nail et al.1976a, b, c; Rousseaux et al. 2002) and including all nanometre-sized amorphousoxides, hydroxides and oxyhydroxides.

• Poorly crystalline pseudo-spinel (Bradley and Hanna 1994) and pseudo-boehmitestructures formed through ageing of the amorphous phases mentioned aboveand containing up to 30 wt% water with respect to the Al2O3 stoichiometry(Tettenhorst and Hofmann 1980; Wefers and Misra 1987; Brinker and Scherer1990). Further ageing of pseudo-boehmite leads to either trihydroxide(s) (atT < �75 ıC) or oxyhydroxide(s) (T > 75ıC) (Wefers and Misra 1987).

• Crystalline Al-trihydroxides (Al(OH)3), in which the Al3C is octahedrallycoordinated by OH� groups, and each OH� is bound to two cations with a vacantthird octahedron, thus forming neutral sheet structures (Demichelis et al. 2009).The most studied Al(OH)3 phases are:

– Gibbsite, the most common and thermodynamically stable phase amonghydroxides in natural systems (major component in bauxite ore). Structurallyit contains OH� ions in consequent and opposite layers with a sequencearranged as f�AB-BA-AB-BA- : : : g in the direction perpendicular to thelayers – it means that OH� ions of the adjacent groups are located directlyin front of each other (Fig. 15.4a) (Saalfeld and Wedde 1974).

– Bayerite, which crystallises under highly basic conditions during processingof bauxites (i.e. the Bayer process). Its layers follow the sequence f�AB-AB-: : : g – OH� ions from one layer are shifted in respect to the previous one, sothat the ions from one layer fall into spaces in-between hydroxyls from theneighbouring ones (Fig. 15.4a) (Rothbauer et al. 1967).

– Nordstrandite, which crystallises upon ageing of precipitates at mildly basicconditions (pH � 7.5–9). Its structure follows the bayerite layer pattern, butwith oppositely located hydroxyl groups, hence yielding the sequence f�AB-AB-BA-BA- : : : g(Fig. 15.4a) (Saalfeld and Jarchow 1968).

– Doyleite, only discovered recently (Caho et al. 1985), has a similar layerstacking as bayerite (Fig. 15.4a), but the hydroxyls of the two consequentdouble layers are located in intermediate positions between those of gibbsiteand nordstrandite (Demichelis et al. 2009).

• Al-oxyhydroxides (AlOOH) are composed of double layers, in which the primarymotif 2 AlOOH forms chains (Wefers and Misra 1987):

– Boehmite, precipitating upon neutralisation of aluminate solutions at elevatedtemperatures, in which 2 AlOOH-based chains are arranged in a cubic packing(Fig. 15.4a) (Christoph et al. 1979).

– Diaspore, in which the 2 AlOOH-based chains are arranged in a hexagonalpacking (Fig. 15.4a) (Busing and Levy 1958). This phase is typically found asthe product of weathering or high-temperature transformations (Keller 1978).

Fig. 15.4 (a) Structures of aluminium oxy-hydroxide phases in octahedral representations. (b)Simulated diffraction patterns of the same phases for the wavelength corresponding to Cu K’1.Patterns and visualisations were generated from the structural information contained in thefollowing entries of the ICSD: gibbsite, 98-000-6162; bayerite, 98-002-6830; nordstrandite, 98-003-4393; doyleite, 98-005-0581; boehmite, 98-003-6340; diaspore, 98-001-6768. Visualisationsand diffraction pattern simulations were prepared in VESTA 3 (Momma and Izumi 2011)


These differences are translated into variations in the diffraction patternsof the various crystalline phases (Fig. 15.4b), which show common structuralfeatures among the trihydroxide group phases in more stark contrast to theoxyhydroxides.

The Al speciation diagram in Fig. 15.1b clearly shows the precipitation regionfor solid trihydroxides at pH �4 to �10. Precipitation has been shown to occurduring the neutralisation of the acidic and alkaline aluminate solutions and pro-ceeds through a colloidal stage with possible polynuclear complex ions and theircondensation products, as will be explained in more detail in the sections below.In both cases, an initially formed amorphous alumina gel eventually transformsto pseudo-boehmite, which has fibrous and sheetlike morphologies. This pseudo-boehmite only gradually crystallises to bayerite and gibbsite (Cesteros et al. 1999)through the following sequence:

Colloidal amorphous or gelatinous alumina .s/ ! pseudo-spinel and pseudo-boehmite ! bayerite ! gibbsite

The polymorphic sequence as well as the morphology and the size distribution ofthe products depend strongly on the physicochemical conditions of the precipitationand transformation reactions (Li et al. 2005a, b, 2011). Gibbsite and bayerite cancrystallise concurrently from aluminate solutions, yet this transformation dependson chemical conditions and bayerite is also known to transform to gibbsite (Lohet al. 2000; Li et al. 2005b, 2011). In strongly caustic environments and at tem-peratures close to 100 ıC, boehmite can form instead of trihydroxides (gibbsite orbayerite) (Wefers and Misra 1987; Gong et al. 2003). The kinetics and mechanismsfor these polymorphic transformations between the various Al phases are still notfully understood, partly because we lack knowledge of their stability field. Indeed,gibbsite is regarded as the most stable phase because it is most common, but clearevidence is missing. For more details, the reader is referred to the excellent reviewsand summaries found in Wefers and Misra (1987), Loh et al. (2000), Gong et al.(2003), Li et al. (2005a) or Li et al. (2011).

15.3.1.1 Alumina Chemistry Under Acidic Conditions

The mechanism and products of acidic hydrolysis of Al3C have been extensivelystudied using a wide variety of in situ and ex situ analytical techniques. Theseinclude potentiometric titrations (Brosset et al. 1954; Nail et al. 1976a), 27Al NMR(Akitt et al. 1988; Akitt 1989; Casey 2006), SAXS (Rausch and Bale 1964; Botteroet al. 1982) and X-ray diffraction (Johansson 1960; Hsu and Bates 1964; Rowselland Nazar 2000).

The process starts at pH < 3 with an unhydrolysed, mononuclear [Al(H2O)6]3C

species, which with increasing pH hydrolyses according to the following idealisedequations (Brinker and Scherer 1990):


ŒAl.H2O/6�3C C xH2O ! ŒAl.OH/x.H2O/6�x�

.3�x/C C xH3OC

where x defines the extent of hydrolysis (i.e. the OH/Al ratio ranging from 0 to 3).The further polycondensation of the so formed ŒAl.OH/x.H2O/6�x�

.3�x/C speciesand how these relate to the final solid Al-polymorphs have been extensively studied,yet the mechanisms are still controversial (Bi et al. 2004 and references therein).Essentially, two condensation mechanisms have been proposed: the “core-links” andthe “cagelike” models (Bi et al. 2004).

According to the “core-links” model, the condensation of hexamericŒAl6.OH/12.H2O/12�

6C rings or ŒAl10.OH/22.H2O/16�8C double rings leads to

larger sheetlike structures. In this case, ŒAl54.OH/144�18C was proposed to be the

last species in the series, before the precipitation of a solid colloidal entity of the[Al(OH)3]n type takes place (Bi et al. 2004). These colloids are expected to exhibita local sheet-structure resembling that of the crystalline Al(OH)3 phases. Thismodel is primarily based on potentiometric titrations (Brosset et al. 1954; Stol et al.1976) combined with X-ray diffraction analyses (Hsu and Bates 1964). Althoughthe “core-links” model explains well the various titration characteristics, it lackssufficient structural evidence for the proposed hexameric units and their consequentcondensation products.

On the other hand, the “cagelike” model considers the well-characterised tride-camer Œ.AlO4/Al12.OH/24.H2O/12�

7C complex ion (Johansson 1960; Johanssonet al. 1960) and other similar structures (Casey 2006) to be the actual primarybuilding units in the alumina system. These primary structures are thought to beinvolved in the nucleation and growth of amorphous colloidal particles and theirtransformation to various crystalline polymorphs (Bottero et al. 1987; Fu et al.1991). The tridecamer unit is also known as the Keggin Al13 structure, and it hasbeen shown that such units exhibit a very high stability (up to several years) (Casey2006). The presence of the Keggin ion in aluminate solutions can be easily identifiedthrough 27Al NMR, and this method has thus been the primary tool to study thegrowth mechanisms and structures involving Keggin complex ions (Akitt 1989).Similarly, the structure of another large Al30 cation, ŒAl30O8.OH/56.H2O/26�

18C,composed of two •-Keggin units linked via four AlO6 units (Fig. 15.5) has alsobeen revealed by 27Al NMR (Rowsell and Nazar 2000; Allouche et al. 2000; Casey2006).

It is important to emphasise that depending on reaction conditions, the formationof other polynuclear species has been predicted. However, due to physical limi-tations of analytical tools (insufficient specificity allowing to distinguish variousAlO6-based species) such as the aforementioned 27Al NMR, many of these specieshave so far not been characterised or even identified (Casey 2006). Hence, resultsfrom the experiments conducted under similar conditions are often interpreted byusing either the “core-links” or the “cagelike” model (see the next paragraph). Therehave been attempts to consolidate the “core-links” and the “cagelike” models byreinterpreting the various experimental data and creating a “continuous” model (Biet al. 2004). This unified model assumes that the polynuclear species transform


Fig. 15.5ŒAl30O8.OH/56.H2O/26�

18C

(aq) (Al30) complex ionshown in a polyhedralexploded view. This structurecan be viewed as two Kegginunits (Al13) joined by fourAlO6 octahedra. Al13 andAl30 species are 1 and 2 nm insize, respectively (Printedwith permission from Casey(2006). Copyright AmericanChemical Society 2006)

rotated trimeric groupδ−Al13moieties

with corner-shared AI(O)6

tetrameric cap

molecule

µ3-OH

on a δ−Keggin

dynamically through different intermediate stages, where the long-lived Keggincomplex ions form upon ageing (i.e. the ion is stable as long as pH or temperaturedoes not drastically change), but the validity of this combined model is still debated.

For instance, Nail et al. studied the early stages of amorphous alumina gel for-mation from AlCl3 and Al2(SO4)3 solutions and postulated a mechanism in whichthe evolution of polynuclear species to a solid colloidal precipitate was a stepwiseprocess leading to the condensation of hexamers (Nail et al. 1976a). Dependingon the original counter anion in the Al3C solution, the colloidal precipitates weresuggested to be composed of �10 six-membered rings for Cl� and only �3 forSO4

2�. Leetmaa et al. studied products forming under nearly identical conditions bymeans of 27Al NMR and infrared/Raman spectroscopy (Leetmaa et al. 2014). Yet,their data also revealed that the products were primarily composed of polynuclearchains of highly distorted aluminium-based octahedra for the Cl� system and ofsulphate-stabilised Keggin Al13 clusters for the SO4

2� system. Perry and Shafranevaluated the optimum conditions for the formation of polynuclear Al complexesand found that the neutralisation of AlCl3 and Al(NO3)3 by KOH and KHCO3 at lowionic strengths favours the formation of Al13 clusters and larger polymers (Perry &Shafran 2001). In contrary, in the presence of sulphates, the resulting products wereprimarily Al13 clusters, which was expected due to the stabilisation of the clustersby SO4

2�.The above-cited studies show that we have been learning more and more about

the structures and the synthetic routes for the various polynuclear species foralumina under acidic conditions. Such insight is very important for the systematicanalysis of the process, because the mechanistic details of the interplay of thevarious clusters on the path the crystalline phase remain rather vague at the moment.


15.3.1.2 Alumina Chemistry Under Alkaline Conditions

Processing of bauxite ores (aluminium-rich ores containing primarily gibbsite,boehmite and diaspore) under alkaline conditions is known as the Bayer method,and it is used to form also gibbsite and nordstrandite (Counter et al. 1997). Thespeciation of Al in basic solutions is rather complex and unsurprisingly a subjectof controversy (Sipos 2009). Furthermore, little is known about any polynuclearcomplex ions that may form under these conditions. In fact, under highly alkalineconditions, the aluminate systems are considered to be actual solutions (seeFig. 15.1b), rather than colloidal systems, as was demonstrated, for instance, bylight scattering or infrared and Raman spectroscopies (Li et al. 2003; Sipos 2009).Typically among the aqueous species, only tetrahedral ŒAl.OH/4�

� is expected tobe present at pH > �8–9 (Moolenaar et al. 1970; van Straten et al. 1984; Swaddleet al. 1994; Li et al. 2003; Sipos 2009). Some other mononuclear species suchas linear [AlO2]� and square planar ŒAl.OH/4.H2O/2�

� complexes may also bepresent (Gerson et al. 1996), but at insignificant concentrations (Swaddle et al. 1994;Sipos et al. 1998; Sipos 2009). At pH > 13, Al3C can exist as either ŒAl.OH/5�

2�

or ŒAl.OH/6�3� (Akitt and Gessner 1984; Barcza and Pálfalvi-Rózsahegyi 1989;

Buvári-Barcza et al. 1998), but again these two species are minority species(Sipos 2009) compared to the main ŒAl.OH/4�

� species. Using computationalapproaches, Gale et al. showed that at alkaline conditions dimeric complexes likeŒ.HO/3Al.OH/2Al.OH/3�

2� should also be expected (Gale et al. 1998), whilethe presence (or significance) of higher polynuclear species is still a subject ofcontroversy (an insightful discussion is given in Sipos (2009) and the referencestherein).

It is known that aluminium in amorphous precipitates is octahedrally coordinatedin the same way as it is later in crystalline phases (Tettenhorst and Hofmann 1980;Wefers and Misra 1987). Upon precipitation from acidic solutions, independentof the actual condensation mechanism and possible speciation, one can assumethat Al3C also condenses into solids with a sixfold coordination. However, in theBayer process, the dominant solution species, i.e. the ŒAl.OH/4�

� ion, exhibitsa fourfold coordination. Hence, one might rationalise the nucleation stage interms of a transition product, in which the realignment of ligands would occurto establish the octahedral geometry by a swap from the tetrahedral one. Thequestion arises whether abundant ŒAl.OH/4�

� species constitute a monomer incrystal growth or whether there are actually other minority species which controlthe process. Gerson et al. carried out semiempirical quantum modelling of var-ious mononuclear, dimeric, trimeric and tetrameric species in caustic aluminatesolutions and suggested a mechanism and possible pathway in which a minoritytetrahedral [Al(OH)3(H2O)] complex could dimerise and then transform to thetetramer, via a transient trimer moiety (Gerson et al. 1996). In such a tetramer, acentral octahedral Al3C ion would be surrounded by three tetrahedral counterparts.However, results from conductivity studies combined with viscosity measurementsof caustic aluminate solution were interpreted in terms of the formation of the


ŒAl6.OH/24�6� (Barcza and Pálfalvi-Rózsahegyi 1989; Buvári-Barcza et al. 1998)

or ŒAl6.OH/22�4� moieties (Sipos et al. 1998). In such species, octahedral Al3C

geometry dominates, yet actual structural evidence for the proposed mechanismsis still lacking to say the least (Sipos 2009). In more dilute caustic solutions (i.e.ŒNaOH� = ŒAl� � 1; and ŒNaOH� � ŒAl� � 1:0 mol=L), a polynuclear speciesin a size range comparable with that of Keggin Al13-like species was proposedbased on primarily light scattering evidence (Li et al. 2003, 2005a). Similarcolloidal polynuclear aluminate structures were suggested based on optically clearsupersaturated aluminate Bayer liquors analysed by cryo-vitrification of solutionsat different stages and subsequent imaging by transmission electron microscopy(Counter et al. 1999). In these systems, the Keggin Al13-like species were assumedto be primarily six-coordinated, and their existence in the Bayer process couldprovide a conceptual link between the mechanisms of Al(OH)3 precipitation fromthe acidic and alkaline solutions. Finally, the critical nuclei size for gibbsite thatforms as an end product in the Bayer process was calculated to have a radiusof � 1.2 nm (Rossiter et al. 1998), which is in line with the size range of Kegginions and other small polynuclear oligomers (cf. Fig. 15.5). This might suggest thatKeggin ions are sort of a universal “precursor” phase in alumina system regardlessof conditions. This is an interesting hypothesis, but yet it has to be confirmed.

To illustrate the influence of physicochemical conditions and speciation ontothe nucleation of phases precipitating from aluminate solutions, let us considerthe investigations reported by Li et al. concerning the growth of gibbsite andbayerite through the Bayer process (Li et al. 2003, 2005a). The authors studiedprecipitation using dynamic light scattering to derive particle size distribution andX-ray diffraction for phase identification. Their data showed that dilute solutions(Al3C

�� 0:82 M and ŒNaOH� � 1:0 M) yielded bayerite and the concentrated

ones (Al3C

�> 2:05 M and ŒNaOH� > 2:5 M) gibbsite, whereas a dimorphic

product precipitated at intermediate concentrations (0:82 <Al3C

�� 2:05 M and

1:0 < ŒNaOH� � 2:5 M). To explain their results, the authors proposed a growthmechanism that took into account differences in speciation in the original aluminatesolutions and was dependent on concentration. For dilute systems (Fig. 15.6a), theysuggested that the process started with Keggin-like complex ions (1) that furtheraggregated and oligomerised to larger clusters (2); these coalesced to amorphousparticles (3), which then lead to the rapid growth of nuclei (of pseudoboehmite)and/or pseudo-spinel phases (4), which transformed into pseudoboehmite (5) andthen crystallised to a more stable bayerite phase. However, for concentratedsolutions (Fig. 15.6b), the process originally involved small aluminate species suchas ion pairs, dimers and trimers rather than the large Keggin ions (1), followed byformation of a loose Al3C-containing oligomeric network (2), its initial densificationto form a cluster (3), further densification to form a crystalline core (4), densifi-cation of the core and nuclei agglomeration (5) and final formation of a gibbsitecrystal (6).


(1) < 5 nm (2) < 20 nm (3) < 200 nm (4) < 200 nm (5) < 200 nm (6) > 200 nm

(1) < 1 nm (2) < 10 nm (3) < 30 nm (4) < 50 nm (5) < 100 nm (6) > 150 nm

A

B

Fig. 15.6 A schematic representation of the formation of crystalline aluminium hydroxide fromsodium aluminate solutions: (a) dilute solutions and (b) concentrated solutions. The size rangesindicated are speculative (Adapted and printed with permission from Li et al. (2005a). CopyrightElsevier 2005)

15.4 Outlook

This chapter highlights the extensive research carried out to unravel the mechanismsand pathways of silica and alumina colloid formation, with regard to both naturalprocesses and industrial applications. Substantial advances in in situ scattering andimaging techniques have allowed to characterise the early stages of colloidal growthand aggregation; however, we are still facing substantial gaps in the molecular-levelmechanistic understanding of the condensation reactions leading to the nucleationof the primary particles. This is particularly true for the alumina system whereanalytical limitations have so far hindered clear identification of all the polynuclearspecies forming in aluminate solutions, making it even harder to define theirrole in the formation of alumina colloids. For the silica system, the presence ofpolynuclear species seems of minor importance and colloid formation is moredependent on the behaviour of the monomeric silica. Whether silica nanoparticlegrowth occurs via the aggregation of primary particles or the addition of monomericsilica to existing particles is however still debated. It is evident that before wecan formulate a fully coherent picture of silica and alumina colloid formation,more studies need to focus on the molecular characterisation of the condensationreactions and the nucleation of the primary particles. This is challenging with thecurrent analytical limitations and the technical difficulties associated with theseobservations. However, in situ scattering and imaging techniques are constantlyimproving in resolution, with regard to both time and space. Also, the importance ofsilica and alumina nanophases for industrial processes further accelerates progressin this field, suggesting that a major breakthrough may come soon.


References

Ab Rahman I, Padavettan V (2012) Synthesis of silica nanoparticles by Sol-gel: size-dependentproperties, surface modification, and applications in silica-polymer nanocomposites—a review.J Nanomater 2012:1687–4110

Akitt JW (1989) Multinuclear studies of aluminium compounds. Prog Nucl Magn Reson Spectrosc21(1–2):1–149

Akitt JW, Gessner W (1984) Aluminium-27 nuclear magnetic resonance investigations of highlyalkaline aluminate solutions. J Chem Soc Dalton Trans 1984:147–148

Akitt JW, Gessner W, Weinberger M (1988) High-field aluminium-27 nuclear magnetic resonanceinvestigations of sodium aluminate solutions. Magn Reson Chem 25(12):1047–1050

Alexander GB (1954) The polymerisation of monosilicic acid. J Am Chem Soc 76(8):2094–2096Allouche L, Gérardin C, Loiseau T, Férey G, Taulelle F (2000) Al30: a giant aluminum polycation.

Angew Chem Int Ed 39(3):511–514Bagwe RP, Hilliard LR, Tan WH (2006) Surface modification of silica nanoparticles to reduce

aggregation and nonspecific binding. Langmuir 22:4357–4362Bale HD, Schmidt PW (1959) Small angle X-ray scattering from aluminium hydroxide gels. II.

J Chem Phys 31(6):1612–1618Barcza L, Pálfalvi-Rózsahegyi M (1989) The aluminate lye as a system of equilibria. Mater Chem

Phys 21(4):345–356Barisik M, Atalay S, Beskok A, Qian S (2014) Size dependent surface charge properties of silica

nanoparticles. J Phys Chem C 118(4):1836–1842Belton DJ, Deschaume O, Perry CC (2012) An overview of the fundamentals of the chemistry of

silica with relevance to biosilicification and technological advances. FEBS J 279(10):1710–20Benning LG, Phoenix V, Mountain BW (2005) Biosilicification: the role of cyanobacteria in silica

sinter deposition. In: Gadd GM, Semple KT, Lappin-Scott HM (eds) Micro-organisms andearth systems: advances in geomicrobiology, SGM Symposium. Cambridge University Press,Cambridge, pp 131–150

Besselink R, Stawski TM, Castricum HL, ten Elshof JE (2013) Evolution of microstructure inmixed niobia-hybrid silica thin films from sol-gel precursors. J Colloid Interface Sci 404:24–35

Bi S, Wang C, Cao Q, Zhang C (2004) Studies on the mechanism of hydrolysis and polymerizationof aluminum salts in aqueous solution: Correlations between the “Core-links” model and“Cage-like” Keggin-Al 13 model. Coord Chem Rev 248:441–455

Bogush GH, Zukoski CF (1991) Uniform silica particle-precipitation – an aggregation growthmodel. J Colloid Interface Sci 142(1):19–34

Bottero JY, Tchoubar D, Cases JM, Fiessinger F (1982) Investigation of the hydrolysis of aqueoussolutions of aluminum chloride. 2. Nature and structure by small-angle x-ray scattering. J PhysChem 86(18):3667–3673

Bottero JY, Axelos M, Tchoubar D, Cases JM, Fripiat JJ, Fiessinger F (1987) Mechanism offormation of aluminium trihydroxide from Keggin Al13 polymers. J Colloid Interface Sci117(1):47–57

Boukari H, Long GG, Harris MT (2000) Polydispersity during the formation and growth of theStober silica particles from small-angle X-ray scattering measurements. J Colloid Interface Sci229(1):129–139

Bradley SM, Hanna JV (1994) 27Al and 23Na MAS NMR and powder X-ray diffraction studies ofsodium aluminate speciation and the mechanistics of aluminium hydroxide precipitation uponacid hydrolysis. J Am Chem Soc 116:7771–7783

Brinker CJ, Scherer GW (1990) Sol-gel science: the physics and chemistry of sol-gel processing.Academic, London

Brosset C, Biedermann G, Gunnar SL (1954) Studies on the hydrolysis of metal ions. XI. Thealuminium ion, Al3C. Acta Chem Scand 8:1917–1926

Busing WR, Levy HA (1958) A single crystal neutron diffraction study of diaspore, AlO(OH).Acta Crystallogr 11:798–803


Buvári-Barcza Á, Rózsahegyi M, Barcza L (1998) Hydrogen bonded associates in the Bayerprocess (in concentrated aluminate lyes): the mechanism of gibbsite nucleation. J Mater Chem8:451–455

Caho GY, Baker J, Sabina AP, Roberts AC (1985) Doyleite, a new polymorph of Al(OH)3, and itsrelationship to bayerite, gibbsite and nordstrandite. Can Mineral 23:21–28

Carcouet CCMC, van de Put MWP, Mezari B, Magusin PCMM, Laven J, Bomans PHH et al (2014)Nucleation and growth of monodisperse silica nanoparticles. Nano Lett 14(3):1433–1438

Carroll S, Mroczek E, Alai M, Ebert M (1998) Amorphous silica precipitation (60 to 120 degreesC): comparison of laboratory and field rates. Geochim Cosmochim Acta 62(8):1379–1396

Casey WH (2006) Large aqueous aluminium hydroxide molecules. Chem Rev 106(1):1–16Cesteros Y, Salagre P, Medina F, Sueiras JE (1999) Several factors affecting faster rates of gibbsite

formation. Chem Mater 11:123–129Christoph GG, Corbató CE, Hofman DA, Tettenhorst R (1979) The crystal structure of boehmite.

Clay Clay Miner 27:81–86Conrad CF, Icopini GA, Yasuhara H, Bandstra JZ, Brantley SL, Heaney PJ (2007) Modeling

the kinetics of silica nanocolloid formation and precipitation in geologically relevant aqueoussolutions. Geochim Cosmochim Acta 71(3):531–542

Counter J, Gerson A, Ralston J (1997) Caustic aluminate liquors: preparation and characterisationusing static light scattering and in situ X-ray diffraction. Colloids Surf A Physicochem EngAsp 126(2–3):103–112

Counter JA, Addai-Mensah J, Raiston J (1999) The formation of Al(OH)3 crystals fromsupersaturated sodium aluminate solutions revealed by cryovitrification-transmission electronmicroscopy. Colloids Surf A Physicochem Eng Asp 154(3):389–398

De Yoreo JJ, Sommerdijk NAJM, Dove PM (2017) Nucleation pathways in electrolyte solutions.In: Van Driessche AES, Kellermeier M, Benning LG, Gebauer D (eds) New perspectives onmineral nucleation and growth, Springer, Cham, pp 1–24

Demichelis R, Catti M, Dovesi R (2009) Structure and stability of the Al(OH)3 polymorphsdoyleite and nordstrandite: a quantum mechanical ab initio study with the CRYSTAL06 code.J Phys Chem C 113:6785–6791

Dietzel M (2000) Dissolution of silicates and the stability of polysilicic acid. Geochim CosmochimActa 64(19):3275–3281

Fouilloux S, Tache O, Spalla O, Thill A (2011) Nucleation of silica nanoparticles measured in situduring controlled supersaturation increase: restructuring toward a monodisperse nonsphericalshape. Langmuir 27(20):12304–12311

Fu G, Nazar LF, Bain AD (1991) Ageing processes of alumina sol-gels: characterization of newaluminium polyoxycations by 27Al NMR spectroscopy. Chem Mater 3:602–610

Gale JD, Rohl AL, Watling HR, Parkinson GM (1998) Theoretical investigation of the natureof aluminum-containing species present in alkaline solution. J Phys Chem B 102(50):10372–10382

Gerson AR, Raiston J, Smart RSC (1996) An investigation of the mechanism of gibbsite nucleationusing molecular modelling. Colloids Surf A Physicochem Eng Asp 110(1):105–117

Gong X, Nie Z, Qian M, Liu J, Pederson LA, Hobbs DT, McDuffle NG (2003) Gibbsite to boehmitetransformation in strongly caustic and nitrate environments. Ind Eng Chem Res 42:2163–2170

Green DL, Lin JS, Lam YF, Hu MZC, Schaefer DW, Harris MT (2003) Size, volume fraction, andnucleation of Stober silica nanoparticles. J Colloid Interface Sci 266(2):346–358

Gunnarsson I, Arnorsson S (2000) Amorphous silica solubility and the thermodynamic propertiesof H4SiO4 degrees in the range of 0 degrees to 350 degrees C at P-sat. Geochim CosmochimActa 64(13):2295–2307

Gunnarsson I, Arnorsson S (2005) Impact of silica scaling on the efficiency of heat extraction fromhigh-temperature geothermal fluids. Geothermics 34(3):320–329

Heaney PJ, Prewitt CT, Gibbs GV (1994) Silica: Physical behaviour, geochemistry & materialsapplications. Reviews in mineralogy Vol 29, Princeton, pp. 606

Hench LL, West JK (1990) The sol-gel process. Chem Rev 90:33–72


Herdianita NR, Browne PRL, Rodgers KA, Campbell KA (2000) Mineralogical and texturalchanges accompanying ageing of silica sinter. Miner Deposita 35(1):48–62

Hsu PH, Bates TF (1964) Formation of X-ray amorphous and crystalline aluminium hydroxides.Mineral Mag 33:749–768

Hsu H-W, Postberg F, Sekine Y, Shibuya T, Kempf S, Horányi M, Juhász A, Altobelli N, SuzukiK, Masaki Y, Kuwatani T, Tachibana S, Sirono S, Moragas-Klostermeyer G, Srama R (2015)Ongoing hydrothermal activities within Enceladus. Nature 519:207–210

Hu C, Liu H, Qu J, Wang D, Ru J (2006) Coagulation behavior of aluminium salts in eutrophicwater: significance of Al13 species and pH control. Environ Sci Technol 40(1):325–331

Icopini GA, Brantley SL, Heaney PJ (2005) Kinetics of silica oligomerization and nanocolloidformation as a function of pH and ionic strength at 25 C. Geochim Cosmochim Acta 69:293–303

Iler RK (1979) The colloid chemistry of silica and silicates. Cornell University Press, IthacaJohansson G (1960) On the crystal structure of some basic aluminium salts. Acta Chem Scand

14:771–773Johansson G, Lundgren G, Sillen LG, Soderquist R (1960) The crystal structure of a basic

aluminum sulfate and the corresponding selenite. Acta Chem Scand 14:769–771Kasprzyk-Hordern B (2004) Chemistry of alumina, reactions in aqueous solution and its applica-

tion in water treatment. Adv Colloid Interface Sci 110:19–48Keller WD (1978) Diaspore recrystallized at low temperature. Am Mineral 63:326–329Kley M, Kempter A, Boyko V, Huber K (2014) Mechanistic studies of silica polymerization

from supersaturated aqueous solutions by means of time-resolved light scattering. Langmuir30(42):12664–12674

Kloprogge JT, Duong LV, Wood BJ, Frost RL (2006) XPS study of the major minerals in bauxite:gibbsite, bayerite and (pseudo-) boehmite. J Colloid Interface Sci 296:572–576

Kobayashi M, Juillerat F, Galletto P, Bowen P, Borkovec M (2005) Aggregation and charging ofcolloidal silica particles: effect of particle size. Langmuir 21(13):5761–5769

Leetmaa K, Gomez MA, Becze L, Guo F, Demopoulos GP (2014) Comparative molecularcharacterization of aluminum hydroxy-gels derived from chloride and sulphate salts. J ChemTechnol Biotechnol 89(2):206–213

Levin I, Brandon D (1998) Metastable alumina polymorphs: crystal structure and transitionsequences. J Am Ceram Soc 81(8):1995–2012

Li H, Addai-Mensah J, Thomas JC, Gerson AR (2003) A study of colloidal Al(III)-containingspecies in fresh/caustic aluminate solutions. Colloids Surf A Physicochem Eng Asp 223:83–94

Li H, Addai-Mensah J, Thomas JC, Gerson AR (2005a) The crystallization mechanism of Al(OH)3from sodium aluminate solutions. J Cryst Growth 279:508–520

Li H, Addai-Mensah J, Thomas JC, Gerson AR (2005b) The influence of Al(III) supersaturationand NaOH concentration on the rate of crystallization of Al(OH)3 precursor particles fromsodium aluminate solutions. J Colloid Interface Sci 286:511–519

Li Y, Zhang Y, Chen F, Yang C, Zhang Y (2011) Polymorphic transformation of aluminum hydrox-ide precipitated from reactive NaAl(OH)4-NaHCO3 solution. Cryst Growth Des 11:1208–1214

Loh JSC, Fogg AM, Watling HR, Parkinson GM, O’Hare D (2000) A kinetic investigation ofgibbsite precipitation using in situ time resolved energy dispersive X-ray diffraction. PhysChem Chem Phys 2:3597–3604

Marshall WL, Chen CTA (1982) Amorphous silica solubilities. 5. Predictions of solubilitybehaviour in aqueous electrolyte solutions of 300ıC. Geochim Cosmochim Acta 46(2):289–291

Matsoukas T, Gulari E (1989) Monomer-addition growth with a slow initiation step: a growthmodel for silica particles from alkoxides. J Colloid Interface Sci 132(1):13–21

Meier D, Gunnlaugsson E, Gunnarsson I, Jamtveit B, Peacock CL, Benning LG (2014) Microstruc-tural and chemical variations in silica-rich precipitates at the Hellisheidi geothermal powerplant. Min Mag 78(6):1381–1389

Momma K, Izumi F (2011) VESTA 3 for three-dimensional visualization of crystal, volumetricand morphology data. J Appl Crystallogr 44:1272–1276


Moolenaar RJ, Evans JC, McKeever LD (1970) Structure of the aluminate ion in solutions at highpH. J Phys Chem 74(20):3629–3636

Mountain BW, Benning LG, Boerema JA (2003) Experimental studies on New Zealand hot springsinters: rates of growth and textural development. Can J Earth Sci 40(11):1643–1667

Nail SL, White JL, Hem SL (1976a) Structure of aluminium hydroxide gel I: initial precipitate.J Pharm Sci 65(8):1189–1191

Nail SL, White JL, Hem SL (1976b) Structure of aluminium hydroxide gel II: aging mechanism.J Pharm Sci 65(8):1192–1195

Nail SL, White JL, Hem SL (1976c) Structure of aluminium hydroxide gel III: mechanism ofstabilization by sorbitol. J Pharm Sci 65(8):1195–1198

Noguera C, Fritz B, Clement A (2015) Precipitation mechanism of amorphous silica nanoparticles:a simulation approach. J Colloid Interface Sci 448:553–563

Pancost RD, Pressley S, Coleman JM, Benning LG, Mountain BW (2005) Lipid biomolecules insilica sinters: indicators of microbial biodiversity. Environ Microbiol 7(1):66–77

Panias D, Asimidis P, Paspaliaris I (2001) Solubility of boehmite in concentrated sodium hydroxidesolutions: model development and assessment. Hydrometallurgy 59:15–29

Patwardhan SV (2011) Biomimetic and bioinspired silica: recent developments and applications.Chem Commun 47(27):7567–7582

Perry CC, Keeling-Tucker T (2000) Biosilicification: the role of the organic matrix in structurecontrol. J Biol Inorg Chem 5(5):537–550

Perry CC, Shafran KL (2001) The systematic study of aluminium speciation in medium concen-trated aqueous solutions. J Inorg Biochem 87:115–124

Petz JJI (1968) Structure of aluminium hydroxide gel. J Chem Phys 48(2):909–911Rausch WV, Bale HD (1964) Small-angle X-ray scattering from hydrolyzed aluminium nitrate

solutions. J Chem Phys 40:3391–3394Rossiter DS, Fawell PD, Ilievski D, Parkinson GM (1998) Investigation of the unseeded nucleation

of gibbsite, Al(OH)3, from synthetic bayer liquors. J Cryst Growth 191:525–536Rothbauer R, Zigan F, O’Daniel H (1967) Verfeinerung der struktur des bayerits, Al(OH)3.

Z Kristallogr Krist 125:317–331Rothbaum HP, Rohde AG (1979) Kinetics of silica polymerization and deposition from dilute-

solutions between 5ıC and 180ıC. J Colloid Interface Sci 71(3):533–559Rousseaux JM, Weisbecker P, Muhr H, Plasari E (2002) Aging of precipitated amorphous alumina

gels. Ind Eng Chem Res 41(24):6059–6069Rowsell J, Nazar LF (2000) Speciation and thermal transformation in alumina sols: structures of

the polyhydroxyoxoaluminum cluster [Al30O8(OH)56(H2O)26]18C and its •-Keggin moieté.J Am Chem Soc 122(15):3777–3778

Ruan HD, Frost RL, Kloprogge JT (2001) Comparison of Raman spectra in characterizing gibbsite,bayerite, diaspore and boehmite. J Raman Spectrosc 32:745–750

Ryan JN, Elimelech M (1996) Colloid mobilization and transport in groundwater. Colloids Surf APhysicochem Eng Asp 107:1–56

Saalfeld H, Jarchow O (1968) Die Kristallstruktur von Nordstrandit, Al(OH)3. Neues JahrbMineral Abh 109:185–191

Saalfeld H, Wedde M (1974) Refinement of the crystal structure of gibbsite, Al(OH)3. Z KristallogrKrist 139:129–135

Sipos P (2009) The structure of Al(III) in strongly alkaline solutions – a review. J Mol Liq 146:1–14

Sipos P, Capewell SG, May PM, Hetter G, Laurenczy G, Lukács F, Roulet R (1998) Spectroscopicstudies of the chemical speciation in concentrated alkaline solutions. J Chem Soc Dalton Trans1998:3007–3012

Smith SJ, Amin S, Woodfield BF, Boerio-Goates J, Campbell BJ (2013) Phase progression of”-Al2O3 nanoparticles synthesized in a solvent deficient environment. Inorg Chem 52:4411–4423

Squyres SW, Arvidson RE, Ruff S, Gellert R, Morris RV, Ming DW et al (2008) Detection ofsilica-rich deposits on Mars. Science 320(5879):1063–1067


Stawski TM, Benning LG (2013) Chapter 5: SAXS in inorganic and bioinspired research. MethodsEnzymol 532:95–127

Stöber W, Fink A, Bohn E (1968) Controlled growth of monodisperse silica spheres in micron sizerange. J Colloid Interface Sci 26(1):62–69

Stol RJ, van Helden AK, de Bruyn PL (1976) Hydrolysis-precipitation studies of aluminum (III)solutions. 2. A kinetic study and model. J Colloid Interface Sci 57(1):115–131

Swaddle TW, Salerno J, Tregloan PA (1994) Aqueous aluminates, silicates, and aluminosilicates.Chem Soc Rev 23:319–325

Tettenhorst R, Hofmann DA (1980) Crystal chemistry of boehmite. Clay Clay Miner 28(5):373–380

Tobler DJ, Benning LG (2013) In situ and time resolved nucleation and growth of silicananoparticles forming under simulated geothermal conditions. Geochim Cosmochim Acta114:156–168

Tobler DJ, Stefansson A, Benning LG (2008) In-situ grown silica sinters in Icelandic geothermalareas. Geobiology 6(5):481–502

Tobler DJ, Shaw S, Benning LG (2009) Quantification of initial steps of nucleation and growth ofsilica nanoparticles: an in-situ SAXS and DLS study. Geochim Cosmochim Acta 73(18):5377–5393

van Straten HA, Holtkamp BTW, de Bruyn PL (1984) Precipitation from supersaturated aluminatesolutions: I. Nucleation and growth of solid phases at room temperature. J Colloid Interface Sci98(2):342–362

Wang L, Zhao W, Tan W (2008) Bioconjugated silica nanoparticles: development and applications.Nano Res 1:99–115

Wefers K, Misra C (1987) Oxides and hydroxides of aluminium, Alcoa Technical Paper 19.Aluminium Company of America, Pittsburgh

Westall F, Walsh MM (2000) The diversity of fossil microorganisms in Archaen-age rocks. In:Seckbach J (ed) Journey to diverse microbial worlds. Kluwer, Amsterdam

Williams LA, Crerar DA (1985) Silica diagenesis 2: general mechanisms. J Sediment Petrol55(3):312–321

Yokoi T, Sakamoto Y, Terasaki O, Kubota Y, Okubo T, Tatsumi T (2006) Periodic arrangement ofsilica nanospheres assisted by amino acids. J Am Chem Soc 128(42):13664–13665

Documents

Chapter 15 Silica and Alumina Nanophases: Natural Processes …homepages.see.leeds.ac.uk/~earlgb/Publications/Tobler et al New... · Silica and Alumina Nanophases: Natural Processes